|
--- |
|
library_name: transformers |
|
license: apache-2.0 |
|
base_model: PrimeIntellect/INTELLECT-1-Instruct |
|
tags: |
|
- axolotl |
|
- generated_from_trainer |
|
datasets: |
|
- neginashz/rationale-llama-chat-dataset |
|
model-index: |
|
- name: star-sft-intellect-instruct-6 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) |
|
<details><summary>See axolotl config</summary> |
|
|
|
axolotl version: `0.6.0` |
|
```yaml |
|
base_model: PrimeIntellect/INTELLECT-1-Instruct |
|
trust_remote_code: true |
|
model_type: AutoModelForCausalLM |
|
tokenizer_config: meta-llama/Llama-3.1-8B-Instruct |
|
|
|
#model_type: LlamaForCausalLM |
|
#tokenizer_type: llama3 |
|
|
|
gpu_memory_limit: |
|
|
|
deepspeed: deepspeed_configs/zero2.json |
|
|
|
load_in_8bit: |
|
load_in_4bit: |
|
strict: false |
|
|
|
chat_template: llama3 |
|
datasets: |
|
- path: neginashz/rationale-llama-chat-dataset |
|
type: chat_template |
|
chat_template: llama3 |
|
field_messages: messages |
|
message_field_role: role |
|
message_field_content: content |
|
roles: |
|
system: |
|
- system |
|
user: |
|
- user |
|
assistant: |
|
- assistant |
|
#roles_to_train: ["assistant"] # default |
|
# Optional[str]. Which EOS tokens to train on in the conversation. Possible values are: |
|
# - all: train on all EOS tokens |
|
# - turn (default): train on the EOS token at the end of each trainable turn |
|
# - last: train on the last EOS token in the conversation |
|
#train_on_eos: turn |
|
|
|
|
|
dataset_prepared_path: |
|
val_set_size: 0.05 |
|
output_dir: ./star-sft-intellect-6 |
|
|
|
sequence_len: 8192 |
|
sample_packing: true |
|
eval_sample_packing: true |
|
pad_to_sequence_len: true |
|
|
|
|
|
wandb_project: star-sft-intellect-instruct-6 |
|
wandb_entity: |
|
wandb_watch: |
|
wandb_name: |
|
wandb_log_model: |
|
|
|
gradient_checkpointing: true |
|
#gradient_clipping: true |
|
gradient_accumulation_steps: 1 |
|
#batch_size: 1 |
|
micro_batch_size: 1 |
|
|
|
num_epochs: 1 |
|
|
|
optimizer: adamw_torch |
|
lr_scheduler: cosine |
|
learning_rate: 0.00002 |
|
|
|
train_on_inputs: false |
|
group_by_length: false |
|
|
|
bf16: true |
|
fp16: false |
|
tf32: false |
|
|
|
logging_steps: 1 |
|
xformers_attention: |
|
flash_attention: true |
|
|
|
warmup_steps: |
|
eval_steps: |
|
save_steps: |
|
|
|
evals_per_epoch: 8 |
|
saves_per_epoch: 2 |
|
eval_max_new_tokens: 128 |
|
|
|
debug: |
|
|
|
weight_decay: |
|
fsdp: |
|
fsdp_config: |
|
|
|
hub_model_id: neginashz/star-sft-intellect-instruct-6 |
|
hub_strategy: |
|
early_stopping_patience: |
|
|
|
resume_from_checkpoint: |
|
auto_resume_from_checkpoints: false |
|
|
|
special_tokens: |
|
pad_token: <|finetune_right_pad_id|> |
|
eos_token": <|eot_id|> |
|
|
|
``` |
|
|
|
</details><br> |
|
|
|
# star-sft-intellect-instruct-6 |
|
|
|
This model is a fine-tuned version of [PrimeIntellect/INTELLECT-1-Instruct](https://huggingface.co./PrimeIntellect/INTELLECT-1-Instruct) on the neginashz/rationale-llama-chat-dataset dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3380 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 2e-05 |
|
- train_batch_size: 1 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- distributed_type: multi-GPU |
|
- num_devices: 4 |
|
- total_train_batch_size: 4 |
|
- total_eval_batch_size: 4 |
|
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 3 |
|
- num_epochs: 1 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 0.4428 | 0.1261 | 14 | 0.4024 | |
|
| 0.433 | 0.2523 | 28 | 0.3939 | |
|
| 0.4197 | 0.3784 | 42 | 0.3799 | |
|
| 0.4083 | 0.5045 | 56 | 0.3679 | |
|
| 0.357 | 0.6306 | 70 | 0.3534 | |
|
| 0.3623 | 0.7568 | 84 | 0.3435 | |
|
| 0.3645 | 0.8829 | 98 | 0.3380 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.47.1 |
|
- Pytorch 2.5.1+cu124 |
|
- Datasets 3.1.0 |
|
- Tokenizers 0.21.0 |
|
|
